Jump to content
Sign in to follow this  
cmunson

Dual Core Doldrums

Recommended Posts

he page cannot be found

The page you are looking for might have been removed, had its name changed, or is temporarily unavailable. Check the name of the page to be sure you have spelled it correctly.

 

Still can't find what you wanted? Please try one of the links shown at right, browse the site navigation on the left, or visit our home page.

something wrong?

Share this post


Link to post
Share on other sites

there are operating systems that take much better advantage of the extra core(s) than windows does. your software hasn't caught up with the technology.

 

amd does have a driver that helps to allieviate this some, and it works well for me.

Share this post


Link to post
Share on other sites
chengrob   

I think that there are some applications that will use dual core more than others. But I still the biggest issue is the user. By nature, I do things serially. Rare are the times, that I am downloading something and then writing a report. Most of the time, I am only doing one thing at a time.

 

Remember, a long time, they came up with hyperthreading, and the sad reality is that the software never caught up to that technology.

 

What I suspect has happened was that Intel/AMD began to run into limitations in the race to continually increase clock speed. Therefore, they starting add another CPU instead of continuing the clock speed race. But the benefit is not even close to as large as when clock speeds were substantially increasing every year.

Share this post


Link to post
Share on other sites
brandon   

I think that there are some applications that will use dual core more than others. But I still the biggest issue is the user. By nature, I do things serially. Rare are the times, that I am downloading something and then writing a report. Most of the time, I am only doing one thing at a time.

 

Remember, a long time, they came up with hyperthreading, and the sad reality is that the software never caught up to that technology.

 

What I suspect has happened was that Intel/AMD began to run into limitations in the race to continually increase clock speed. Therefore, they starting add another CPU instead of continuing the clock speed race. But the benefit is not even close to as large as when clock speeds were substantially increasing every year.

 

What ever happened to the 10Ghz CPUs Intel was supposed to have by 2003? You can only increase clockspeeds so far before you run into a thermal wall. Remember why there was never a 4Ghz Prescott?

 

The faster you run a chip, the lower the yields. You can have a superfast CPU with poor yields, or a slower multicore CPU with better yields. It's a win-win situation. Not only do you get double the amount of cores, and in some cases, almost double the performance of a single core CPU, you can get good yields due to them running slow, since you can rebadge faulty chips as single core offerings.

Share this post


Link to post
Share on other sites

they can do the high clocks and get exceptional yields, many future mark tests prove that, however the oc customers don't really put a dent in sales, they can't imagine every pc shipped with vap0chills :lol:

 

they look at it as running stock when they design these chips, they don't care about gaming or overclocking, they never did. multiple threads, more done in reasonable time, they know they have us :)

Share this post


Link to post
Share on other sites
brandon   

That's not what I mean by yields. I mean the total amount of useable/working chips per wafer. With higher clockspeeds, the risk of a defect becomes exponentially higher.

Edited by brandon

Share this post


Link to post
Share on other sites
Bruce   

Well I have to chime in.

 

The problem you are experiencing Rob has more to do with your operating system then it does your choice of processors.

 

I have an HP dv2037us laptop, with a 1.6 dual core processor, and the thing absolutely flies. I mean it is fast, I was so impressed with the Intel dual core processor in my laptop that I now have two desktops with Intel processors, one with a dual core and one with a quad core.

 

My laptop was formatted the day I bought it and was never booted into windows, I loved it, fast, responsive, handled multiple tasks like no tomorrow and best of all It handled video encoding much faster then a AMD dual core 4200+ desktop machine. I was amazed to say the least.

 

It is an impressive machine.

 

 

A few months after that I was curious how it would run with the OS it shipped with and was initially intended to be used with. That is Windows XP SP2. I was shocked, it was slow, responded like a rowboat with a 1000 pound anchor attached to it. So I started tweaking, and optimizing, and applied every known trick in the book to get it run well. The result? It was slow, responded like a rowboat with a 1000 pound anchor attached to it. Now this is a dual core 1.6 with an sata hard drive and 2 gigabytes of DDR2 memory. I thought, if I had purchased this and used it with XP I would have been seriously disappointed and may have even blamed the hardware, or the manufacturer for it's snail like performance. Opening more then two applications at once and it behaved like a 266 with 8 megs of ram. I simply couldn't believe it.

 

I spent a considerable amount of time and energy into getting it to run well. It just wouldn't.

 

It didn't take me long to go back to my OS of choice. I plug in my WinTV-USB2 television tuner, watch tv, while simultaneously encode video, browse the web, and compile applications from source code. With the exception of the web browsing, all are pretty CPU intensive, and the little laptop just behaves as if I had nothing going on at all. It does it all and does it very well and fast.

 

If you ever want to sell that little 1.2 dual core laptop give me a yell, because I know it can be made to work extremely well. ;)

 

As for the test telling you that it is "slow", well comparing a laptop to all computers will of course say it's slow. I think your tests are not fine grained enough.

 

They should not be comparing laptops with desktops, or workstations with with home computers....................

 

Then take into consideration that when you bought your laptop it was already 2 years old, and you were setting yourself up to have outdated hardware the day you purchased it.

Edited by Bruce

Share this post


Link to post
Share on other sites
chengrob   

Brandon,

 

Back when I was at Gateway, here's how they made processors. Intel actually took me on a tour of their fabrication and the tour guide was Paul Ottelini. I'm showing off my age. We were both in the bunny outfits, too. Anyways, the last phase of the production process is called bin out. This is where they sort the wafers. Despite all of their improvements in manufacturing, not all wafers are identical in performance. So therefore, quality is defined by bin out. Bin out means how they sort the chips based on clock. So if the process ran well, a high percentage of the chips would bin out at the highest clock speed. Because obviously, the higher the clock speed, the more Intel could charge for the chip. Note: we call it quality, but Intel did not throw out any chips. They just would mark them down to lower clock speeds.

 

Our overclocking friends have already figured this out, because frequently Intel would under bin. Let's say that Intel has an exceptionally good run, and 90% binned out at the highest clock speed. Since they cannot sell all the processors at the highest clock speed, they would underbin and actually sell the chips at a lower clock speed. These are the best chips for overclocking. And of course, this is the fodder for the age long debate whether Intel or AMD make the best chips for overclocking.

 

Bruce,

 

I suspect strongly that you are correct. I wonder is there an objective measurement to quantitatively highlight the difference between operating systems.

 

I am definitely disappointed in this computer but I don't think I will sell it. I love the screen, and I like keeping my old computers as memorabilia. I still have a Windows 98 200 Mhz that still runs great. But the screen is 800 x 600.

Share this post


Link to post
Share on other sites
Bruce   

Of course there are. There are a few benchmarks that run on any platform (cross platform), however the little competitions that the various forums set up, won't let Linux or Apple machines compete against them, because simply put we murder them on those tests. They have to "overclock" to get to the same speeds we achieve at stock settings, and if we overclcok then they just loose against the better OS.

 

So their solution to the problem is this.

 

They do not let Linux and Macs to compete against them. ;)

 

So while they claim these are "hardware wars", and "hardware benchmarks" the reality is, they are not, they are "who can keep windows from blue screening contests"

 

If they were truly interested in "hardware wars", and "hardware benchmark" contests then they would allow everyone to participate.

Share this post


Link to post
Share on other sites
brandon   

BB, that is definitely unfair. The driver would be minorly inconvienenced if he were to get poked by the plastic knife. The driver doesn't deserve to be inconvienenced, his job is hard enough already.

Edited by brandon

Share this post


Link to post
Share on other sites
chengrob   

I can understand why the guys don't want to compete against Apple's and Macs.

 

But has there been any objective, empirical research on the differences between operating systems. For example, people are stating that Linux/Apple wins over Windows. By how much. Is the difference big like 50%, or is it more like 5%.

 

Now I am a Vista user for close to a year, and I suspect that Vista is definitely noticeable slower on 2D graphics.

Share this post


Link to post
Share on other sites

 

Now I am a Vista user for close to a year, and I suspect that Vista is definitely noticeable slower on 2D graphics.

 

it has a lot to haul around, huge DRM kernel needs resources and the purdy gui will bring down any graphics embedded chipset.

Share this post


Link to post
Share on other sites
Bruce   

I can understand why the guys don't want to compete against Apple's and Macs.

 

But has there been any objective, empirical research on the differences between operating systems. For example, people are stating that Linux/Apple wins over Windows. By how much. Is the difference big like 50%, or is it more like 5%.

 

Now I am a Vista user for close to a year, and I suspect that Vista is definitely noticeable slower on 2D graphics.

 

That would depend on the particular benchmark Rob. There are lots of different sets of data that show Linux outperforming windows. From disk speed throughput, to tcp/ip throughput, to database speeds, apache speeds and on and on. I think when one looks at the majority of the worlds fastest supercomputers and the OS that runs most of them it becomes obvious why they chose the operating system that they did.

 

In the few tests that they are using in the OC forum that I have run, I would say on average about 20% to 25% faster and sometimes even more. That was with a machine running stock speeds with no tweaks and about 130 background services running :P

 

On a machine I just sold to another member here at the pit. In windows disk speeds were 46 MB/s, in Linux they were 72 MB/s. that is a 36% increase in disk speed, on the same exact hardware.

Share this post


Link to post
Share on other sites

 

 

On a machine I just sold to another member here at the pit. In windows disk speeds were 46 MB/s, in Linux they were 72 MB/s. that is a 36% increase in disk speed, on the same exact hardware.

 

that is so true bruce, on some old machines with windows using fake raid if it matters my disk score was 99 mb's, when running linux on those same machines my disk speed rated 138 mb's :tup:

Share this post


Link to post
Share on other sites
Kenny3   

I agree its the OS that can cripple hardware, this is why I run Kubuntu on my laptop. its only a p3 but when i had XP on it, it use to choke running most applications. as soon as I switched it to a linux distro, it was like a whole new machine. still chugs along nicely. and a pleasure to surf with :tup:

Share this post


Link to post
Share on other sites

I haven't read the whole thread, and I will shortly, but what are the benchmarks that Linux and Windows can both run. I've been looking for something different to use for our little competitons here.

Share this post


Link to post
Share on other sites
Bruce   

Super PI, geekbench, are two that they are using that will run "natively" in Linux.

 

But it is irrelevant, because they will not allow you to use Linux in their competitions at all.

Share this post


Link to post
Share on other sites

in fact if you think about it, the compiler gave the benchmark result.

 

 

Right now I'm aware of a couple of benchmarks that are fast on platforms that use GCC (Mac OS X and Linux) and slow on platforms that use Windows. It's not that the operating system is better (or faster), it's that Visual C++ isn't able to optimize some of the code in Geekbench as well as GCC. This doesn't mean the underlying hardware is slower, it's just that the hardware is running code that isn't optimized as well as it could be.

Edited by duanester

Share this post


Link to post
Share on other sites
tito   

I don't think the compiler can take all the credit. If you compile an "x86" binary or an "x86-64" binary in GCC for Linux and distribute that, it won't be half as optimised as if you compiled it yourself with the right flags for your system.

Share this post


Link to post
Share on other sites

i want to verify this, i am not the compiler king, i compiled a few things from source but thats it and it only took seconds.

 

 

i will wait for more experienced persons patiently :tup:

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this  

×